64
C.B. Elias and J.B. Joshi
precipitates obtained in the two different modes of precipitation were further
characterized. Sedimentation of the batch tank prepared material with a large
mean size resulted in a hazy supernatant. In contrast the material from the
tubular reactor resulted in a clear supernatant. The particles obtained from the
tubular reactor were found to be easily disrupted in subsequent capillary shear
experiments and prior aging in a stirred reactor resulted in improved aggregate
strength and recovery. Microscopic observation of the particles prepared by the
two systems showed that the continuously prepared aggregate was irregular in
shape and hence more susceptible to disruption as compared with that prepared
in the stirred tank. The very wide range of velocity gradients existing in the
stirred vessel coupled with increased residence time resulted in improved aggregate strength. This is supported by the observations of Virkar et al. [-66] who
found that use of grid turbulence promoters installed at regular intervals
improved the particle strength in a continuous tubular reactor.
The centrifugal separation of the precipitates was further examined by Bell
and Dunnill [68] using disk and scroll discharge centrifuges. For the disk
centrifuge no significant differences on the separation efficiency were seen for the
precipitate prepared by the different methods. However, in the scroll discharge
or decanter centrifuge, the batch prepared precipitate gave 30% greater
throughput before efficiency dropped below 98 %, while the solid content for the
tubular reactor material was approximately 30% higher than the decanter
centrifuge material. It is possible that the particles produced by shear breakup in
the equipment and resulting in fines are not recovered, i.e. when the throughput
is increased the clarification efficiency is reduced due to non recovery of fines
which are probably produced in the feed zone. The stirred tank prepared
material showed considerable resistance to long exposure times of turbulent
shear, suggesting that disruption in the scroll centrifuges occurs due to the
intense shear in the feed zone and not due to turbulence existing in the bowl of
the centrifuge. This was supported by the results of the short duration high shear
capillary experiments where the sizes of particles generated by breakup were
similar to the unrecovered particles in the scroll centrifuges. This work clearly
points out the need of design of centrifugal separation processes keeping in mind
the nature of precipitates and integrating the precipitation and centrifugal
recovery operations.
Aggregate stability may be described by equating the size dependent hydrodynamic forces to the size dependent strength of the aggregate. The dependence on shear rate takes the form
D,,,x~G -~
(12)
where Dm~x is the largest possible size which can avoid disruption and 7 is
indicative of the mechanism of breakup [69, 70]. The disadvantage of this model
is that there is no information on the kinetics of breakup or on the size of
distribution below the maximum size. Twineham et al. [71] proposed a displacement model describing the aggregation process, including breakup in terms of
the rate of change of aggregate diameter at a fixed cumulative volume percent.
C.B. Elias and J.B. Joshi
precipitates obtained in the two different modes of precipitation were further
characterized. Sedimentation of the batch tank prepared material with a large
mean size resulted in a hazy supernatant. In contrast the material from the
tubular reactor resulted in a clear supernatant. The particles obtained from the
tubular reactor were found to be easily disrupted in subsequent capillary shear
experiments and prior aging in a stirred reactor resulted in improved aggregate
strength and recovery. Microscopic observation of the particles prepared by the
two systems showed that the continuously prepared aggregate was irregular in
shape and hence more susceptible to disruption as compared with that prepared
in the stirred tank. The very wide range of velocity gradients existing in the
stirred vessel coupled with increased residence time resulted in improved aggregate strength. This is supported by the observations of Virkar et al. [-66] who
found that use of grid turbulence promoters installed at regular intervals
improved the particle strength in a continuous tubular reactor.
The centrifugal separation of the precipitates was further examined by Bell
and Dunnill [68] using disk and scroll discharge centrifuges. For the disk
centrifuge no significant differences on the separation efficiency were seen for the
precipitate prepared by the different methods. However, in the scroll discharge
or decanter centrifuge, the batch prepared precipitate gave 30% greater
throughput before efficiency dropped below 98 %, while the solid content for the
tubular reactor material was approximately 30% higher than the decanter
centrifuge material. It is possible that the particles produced by shear breakup in
the equipment and resulting in fines are not recovered, i.e. when the throughput
is increased the clarification efficiency is reduced due to non recovery of fines
which are probably produced in the feed zone. The stirred tank prepared
material showed considerable resistance to long exposure times of turbulent
shear, suggesting that disruption in the scroll centrifuges occurs due to the
intense shear in the feed zone and not due to turbulence existing in the bowl of
the centrifuge. This was supported by the results of the short duration high shear
capillary experiments where the sizes of particles generated by breakup were
similar to the unrecovered particles in the scroll centrifuges. This work clearly
points out the need of design of centrifugal separation processes keeping in mind
the nature of precipitates and integrating the precipitation and centrifugal
recovery operations.
Aggregate stability may be described by equating the size dependent hydrodynamic forces to the size dependent strength of the aggregate. The dependence on shear rate takes the form
D,,,x~G -~
(12)
where Dm~x is the largest possible size which can avoid disruption and 7 is
indicative of the mechanism of breakup [69, 70]. The disadvantage of this model
is that there is no information on the kinetics of breakup or on the size of
distribution below the maximum size. Twineham et al. [71] proposed a displacement model describing the aggregation process, including breakup in terms of
the rate of change of aggregate diameter at a fixed cumulative volume percent.